DOI: https://doi.org/10.14710/j.gauss.11.4.562-571

PENGGUNAAN SELEKSI FITUR CHI-SQUARE DAN ALGORITMA MULTINOMIAL NAÏVE BAYES UNTUK ANALISIS SENTIMEN PELANGGGAN TOKOPEDIA

*Tri Ernayanti  -  Departemen Statistika, Fakultas Sains dan Matematika, Universitas Diponegoro, Indonesia
Mustafid Mustafid  -  Departemen Statistika, Fakultas Sains dan Matematika, Universitas Diponegoro, Indonesia
Agus Rusgiyono  -  Departemen Statistika, Fakultas Sains dan Matematika, Universitas Diponegoro, Indonesia
Arief Rachman Hakim  -  Departemen Statistika, Fakultas Sains dan Matematika, Universitas Diponegoro, Indonesia
Open Access Copyright 2022 Jurnal Gaussian under http://creativecommons.org/licenses/by-nc-sa/4.0.

Citation Format:
Abstract
E-commerce is a medium for online shopping that is popular among the public. Ease of access for all internet users and the completeness of products offered by e-commerce are new alternatives in meeting the needs of the community. This causes stiff competition in the e-commerce, so e-commerce need to carry out the right marketing strategy in order to compete in obtaining, retaining, and partnering with customers, one of which is by reviewing aspects of customer satisfaction. Tokopedia is an e-commerce buying and selling online that connects sellers and buyers throughout Indonesia for free. In this study, an analysis of Tokopedia's customer sentiment was carried out with the Multinomial Naïve Bayes classification. Algorithm Multinomial Nave Bayes is a model development of the Nave Bayes. The difference lies in the selection of data, if Naïve Bayes uses a Gaussian that is suitable for continue, while Multinomial Naïve Bayes is suitable for discrete data such as the number of words in a document. Multinomial Naïve Bayes is the simplest method of probability classification but is sensitive to feature selection, so the amount of data is determined by the results of Chi-Square.Multinomial Naïve Bayes is used to classify customer opinions that are positive and negative so that they can form customer satisfaction factors Tokopedia, while the Chi-Square used to measure the level of feature dependence with class (positive and negative) so as to eliminate disturbing features in the classification process. Classification performance results using Multinomial Naïve Bayes without Chi-Square obtained accuracy and kappa statistics of 88% and 75.95%, while using Chi-Square obtained accuracy and kappa statistics of 95% and 89.99%, respectively. This means that Multinomial Naïve Bayes has quite effective performance and results in analyzing Tokopedia customer satisfaction sentiment and the use of Chi-Square for feature selection can improve the accuracy of the classification process. 
Fulltext View|Download
Keywords: E-commerce; Sentiment Analysis; Multinomial Naϊve Bayes; Feature Selection Chi-Square.

Article Metrics:

Article Info
Section: Articles
Language : ID
Recent articles
PERBANDINGAN SMOTE DAN ADASYN PADA DATA IMBALANCE UNTUK KLASIFIKASI RUMAH TANGGA MISKIN DI KABUPATEN TEMANGGUNG DENGAN ALGORITMA K-NEAREST NEIGHBOR PENGARUH KONVEKSITAS TERHADAP SENSITIVITAS HARGA JUAL DAN DELTA-NORMAL VALUE AT RISK (VAR) PORTOFOLIO OBLIGASI PEMERINTAH MENGGUNAKAN DURASI EKSPONENSIAL PEMODELAN KURS RUPIAH TERHADAP DOLAR AMERIKA SERIKAT MENGGUNAKAN REGRESI NONPARAMETRIK CAMPURAN KERNEL DAN SPLINE More recent articles
  1. APJII. (2020). Laporan Survei Penetrasi dan Profil Perilaku Pengguna Internet Indonesia Tahun 2020. Jakarta: Asosiasi Penyelenggara Jasa Internet Indonesia
  2. Feldman, R., dan Sanger, J. (2007). The Text Mining Handbook: Advanced approaches in Analyzing Unstructured Data. New York: Cambridge University Press
  3. Han, J., dan Kamber, M. (2006). Data Mining: Concept and Techniques (2nd ed.). San Fransisco: Morgan Kaufmann Publisher
  4. Hartanto. (2017). Text Mining dan Sentimen Analisis Twitter pada Gerakan LGBT. Jurnal Psikologi Ilmiah, IX(1), 18-25
  5. Kotler dan Keller. (2014). Manajemen Pemasaran Edisi 13 Jilid 1. Jakarta: Erlangga
  6. Ling, J., Kencana, I. P., dan Oka, T. B. (2014). Analisis Sentimen Menggunakan Metode Naive Bayes Classifier Dengan Seleksi Fitur Chi Square. E-Jurnal Matematika, Vol. 03
  7. Liu, B. (2012). Sentiment Analysis and Opini Mining. San Rafael: Morgan dan Claypool Publishers
  8. Liu, J., Tian, Z., Liu, P., Jiang, J. & Li, Z. (2016). An Approach of Semantic Web Service Classification Based on Naive Bayes. IEEE International Conference on Services Computing, Hal: 356-362
  9. McCallum, A., dan Nigam, K. (1998). A Comparison of Event Models for Naive Bayes Text Classification. Pitssburgh: Proceedings in Workshop on Learning for Text Categorization
  10. Sun, C, X. Wang, and J. Xu. (2009) "Study on Feature Selection in Finance Text Categorization". Science And Technology, pp. 5077-5082
  11. Sung, J., dan Sung, E. (2014). Exploring the Usefulness of a Decision Tree in Predicting People Locations. Procedia - Social and Behavioral Sciences, 140, 447–451
  12. Taheri, S., dan Mammadov, M. (2013). Learning the Naive Bayes Classifier with Optimization Models. International Journal of Applied Mathematics and Computer Science, 787-795
  13. Wati, R. (2016). Penerapan Algoritma Genetika untuk Seleksi Fitur pada Analisis Sentimen Review Jasa Maskapai Penerbangan Menggunakan Naïve Bayes. Evolusi Vol. 4, No. 1, Hal: 25-31
  14. Witten, L., Frank, E., dan Hall, M. A. (2011). Data Mining: Practical Machine Learning Tools and Techniques (3rd ed.). USA: Elsevier

Last update:

No citation recorded.

Last update:

No citation recorded.